Application Research Of Geohazard Risk Assessment Based On SSA Optimized SVM

Due to the influence of geological environment and various inducing factors,there are many environmental geological problems in Dingcheng District of Changde City,Hunan Province.The frequent geohazards directly affect the development of local social economy.The risk assessment of geohazards in the study area can provide reference for the formulation of disaster prevention and mitigation projects and related policies in the study area.In this paper,the slope unit is selected as the evaluation unit,and the support vector machine(SVM)model optimized by sparrow search algorithm(SSA)is used to evaluate the geohazard risk in the study area.On the basis of risk assessment,the risk assessment results of geohazards in Dingcheng District were obtained by superposition with vulnerability assessment results.The main research contents and results are as follows :(1)Take slope unit as evaluation unit.Hydrological analysis method and surface curvature watershed method are used to complete the division of slope units in the study area.In the process of division,the influence of the threshold of flow accumulation value of hydrological analysis method and the elevation filtering value of surface curvature watershed method on the final division results of slope units is discussed.Slope units under different methods and grading values are combined with SSA-SVM model to predict the hazard of geohazards.By comparing the hazard prediction results and ROC curve,it is determined that the slope unit divided when the flow accumulation value is 1000 and the elevation filtering value is 10 has higher accuracy and AUC value.It is determined that the slope unit divided when the flow accumulation value is 1000 and the elevation filtering value is 10 is the evaluation unit of this risk assessment.(2)Determinate hazard assessment indicators and perform sensitivity analysis.Through data collection and analysis,a total of 10 indicators including slope,topographic relief,elevation,slope direction,geological structure,rock and soil type,distance to river system,vegetation coverage(VC),human engineering activities and annual average rainfall were determined as geohazard hazard assessment indicators.Combined with the prediction and zoning results of SSA-SVM model,the sensitivity of each evaluation index was analyzed.It is concluded that the three evaluation indicators of slope,topographic relief and human engineering activities have a great influence on the accuracy of the model and have high sensitivity.The sensitivity of annual rainfall and rock and soil types in geohazard hazard assessment is relatively weak.The sensitivity of distance to river in the geohazard hazard assessment is the weakest.(3)Three vulnerability indexes,such as population density,buildings and roads,were selected to establish the vulnerability evaluation index system of geohazards in Dingcheng District,and the vulnerability zoning of geohazards in the study area was obtained by the method of assignment and superposition.Using the risk assessment model ’ R = H × V ’ for risk superposition,combined with the risk level judgment matrix,the risk assessment of geohazards in the study area was completed.